predict

Generate predictions on new data. Returns class labels (classification) or numeric values (regression) by default. Pass proba=True for class probabilities.

Signature

ml.predict(model, data, *, proba=False, intervals=False, confidence=0.90)

predict(model, newdata, proba = FALSE)

Parameters

Parameter	Type	Default	Description
`model`	Model	—	A fitted model
`data`	DataFrame	—	New data (same features as training)
`proba`	bool	`False`	Return class probabilities instead of labels
`intervals`	bool	`False`	Return prediction intervals (regression, Python only)
`confidence`	float	`0.90`	Confidence level for prediction intervals

Returns

A Series of predictions (labels or values), or a DataFrame when proba=True or intervals=True.

Examples

Class predictions

preds = ml.predict(model, new_data)
print(preds.head())

preds <- predict(model, new_data)
head(preds)

Probabilities

probs = ml.predict(model, new_data, proba=True)
# Returns DataFrame with one column per class

probs <- predict(model, new_data, proba = TRUE)

validate →